Abstract
Soil erosion rates are accelerating worldwide as climate change effects and human population pressures, including agricultural expansion, degrade the land surface. Fluvial systems transfer sediments from uplands to depositional landforms and basins downstream. However, only a fraction of eroded material will ultimately transfer to catchment outlets – a phenomenon termed, “the sediment delivery problem”. Thus, sediment fluxes to the coast are declining in many river catchments, as a result of storage behind dams and within landforms such as floodplains and alluvial terraces. Storage time allows us to measure the timescale of storage and removal of sediments from floodplains, which, given their spatial extent (8 x 105 to 2 x 106 km2 of all land area), are significant in interrupting the transmission of soil erosion fluxes downstream. While sediment storage times in alluvial floodplains have been quantified before, this thesis presents the first attempts to model the impacts of various environmental and experimental conditions on sediment storage behaviour using the CAESAR-Lisflood landscape evolution model. The thesis tests the following hypotheses: i) Removal rates from storage decline with increasing floodplain age; ii) the distribution of sediment storage times is sensitive to reach-specific characteristics, vegetation cover types and changes, changing river flows, and measurement frequency; and iii) a non-linear function can be fitted to the distribution and parameterised using readily quantifiable variables. A detailed literature review synthesised our current understanding of sediment storage times, including variables that have been quantified or hypothesised as possible controls. This culminated in a conceptual model of major controls and their interactions which was used to support the development of experiments tested in this thesis. A review of quantification techniques, including “black-box”, one-dimensional mass balance modelling approaches, and methods that calculate storage times directly from timings of geomorphic changes, justified adopting a landscape evolution modelling approach. CAESAR-Lisflood was applied to conduct this research, as it can simulate variable channel widths, divergent flow, and both braided and meandering planforms – capturing a wider range of channel-floodplain evolution processes than models previously used to simulate storage times. Ten 1 km-long reaches of river valleys from the north of England were used to calibrate the model, test the transferability of calibrated parameters, and verify the accuracy of simulated historical channel changes against mapped reconstructions. These simulations replicated mapped erosion, deposition and lateral migration rates reasonably well overall. Floodplain turnover times, estimated by extrapolating erosion rates, increased confidence that calibrated parameters were representative over longer timescales and revealed that all sediments stored in the floodplain would undergo exchange with the channel within 1000 years. Using CAESAR-Lisflood, an ensemble of 9 simulations, incorporating 3 of the 10 calibrated reaches and 3 vegetation cover scenarios (forest, grass and unvegetated) – each spanning 1000 years of river channel changes – was run. Together with measuring channel changes over four different frequencies (10, 20, 50 and 100 years), a total of 36 storage time distributions was modelled, with the age and storage times of floodplain sediments calculated from timings of deposition and erosion. This was done to test whether distributions were best fit by either an exponential or a heavy-tailed decay function, with the former indicating constant erosion rates over space and time, while the latter implies that removal rates from storage decay with increasing deposit age. As well as uniform vegetation conditions, a further 15 simulations, incorporating changes in vegetation cover or flow magnitudes over time, were run, to test how storage time dynamics respond to disturbance. This thesis demonstrates that sediment erosion rates decline with increasing floodplain age in most cases, with the strength of this relationship dependent on reach, floodplain erodibility and frequency of recorded measurements. A lognormal function can be fitted to distributions of sediment storage times in most cases, and it is possible to parameterise this function using the median storage time and measurement time-step. Coupling this storage time function with a model of stochastic sediment transport could generate predictions of decontamination times for a valley corridor enriched with polluted sediments (e.g. from mining). However, some environmental disturbances can be great enough to invalidate this storage time model – a challenge that merits further attention before application to practical environmental management contexts.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.